This study is a secondary analysis of a state-wide nursing home data set, based on two assessments of nursing home residents (6-month followup) in Rode Island, representing 105 facilities and 3,300 residents. The major purpose of this study is to improve the understanding of the risk factors associated with physical functional change and discharge status. A secondary purpose is to demonstrate the potential applicability of the study results to quality assurance, quality regulation, and prevention strategies. Based on estimation of logit models, dichotomous variables (physical decline, improve, and discharge to community) will be regressed on a series of patient characteristic models, facility characteristic models, and combined models to determine if facility characteristics affect outcomes, after controlling for patient characteristics. Estimates of expected outcomes for types of residents will be developed from these models. In addition, as a case study at the facility level, facilities will be screened based on the relation between actual outcomes and expected outcomes predicted from patient regressions. If facility characteristics are significant, expected outcome estimates will be re-estimated using residents in "better" nursing homes. Implications of results for quality assurance, quality regulation, and prevention will be studied.